Silencing of Reversion-Inducing Cysteine-Rich Protein with Kazal Motifs Stimulates Hyperplastic Phenotypes through Activation of Epidermal Growth Factor Receptor and Hypoxia-Inducible Factor-2α

Reversion-inducing cysteine-rich protein with Kazal motifs (RECK, a tumor suppressor) is down-regulated by the oncogenic signals and hypoxia, but the biological function of RECK in early tumorigenic hyperplastic phenotypes is largely unknown. Knockdown of RECK by small interfering RNA (siRECK) or hypoxia significantly promoted cell proliferation in various normal epithelial cells. Hypoxia as well as knockdown of RECK by siRNA increased the cell cycle progression, the levels of cyclin D1 and c-Myc, and the phosphorylation of Rb protein (p-pRb), but decreased the expression of p21cip1, p27kip1, and p16ink4A. HIF-2α was upregulated by knockdown of RECK, indicating HIF-2α is a downstream target of RECK. As knockdown of RECK induced the activation of epidermal growth factor receptor (EGFR) and treatment of an EGFR kinase inhibitor, gefitinib, suppressed HIF-2α expression induced by the silencing of RECK, we can suggest that the RECK silenicng-EGFR-HIF-2α axis might be a key molecular mechanism to induce hyperplastic phenotype of epithelial cells. It was also found that shRNA of RECK induced larger and more numerous colonies than control cells in an anchorage-independent colony formation assay. Using a xenograft assay, epithelial cells with stably transfected with shRNA of RECK formed a solid mass earlier and larger than those with control cells in nude mice. In conclusion, the suppression of RECK may promote the development of early tumorigenic hyperplastic characteristics in hypoxic stress.


Introduction
Reversion-inducing cysteine-rich protein with Kazal motifs (RECK) is a tumor suppressing, membrane-anchored glycoprotein that contains multiple epidermal growth factor-(EGF)-like repeats and multiple serine protease inhibitor-like domains [1]. RECK inhibits proMMP-2 activation and the enzymatic activities of MMP-2, MMP-9, and membrane type 1 (MT1)-MMP [2,3]. Although RECK functions as an inhibitor of matrix metalloproteinases (MMPs), it does not share structural homology with tissue inhibitors of metalloproteinases (TIMPs) [2]. RECK is expressed in various normal human tissues, but has not been detected in oncogenically transformed cells and in various type of cancers, such as, hepatoma, pancreatic, breast, lung, colorectal, prostate, and gastric cancer, or in osteosarcomas [4]. RECK downregulation in cancer tissues is associated with a low survival rate and a poor prognosis because RECK inhibits angiogenesis, invasion, and metastasis in cancer via MMP inhibition [4]. Germ line knockout of RECK induced the proliferation of mouse embryonic fibroblasts (MEFs) and enabled early escape from the cellular senescence induced by oncogenic insults, and RECK interferes with epidermal growth factor receptor (EGFR) signaling [5]. Epigenetics, such as, histone and DNA modifications, are involved in the silencing of RECK [6,7]. We previously reported that RECK is downregulated under hypoxic conditions through HDAC1 and hypoxia inducible factor (HIF)-1α [8]. Several microRNAs are also involved in targeting RECK in hypoxia and RAS-mediating signaling pathways [9]. Thus, it has been suggested that HIF-1α is a key regulator to inhibit RECK expression in tumorigenesis [8,9].
Hypoxia (a reduction in tissue oxygen tension) is frequently detected in growing tumors larger than 1 mm 3 [10]. Interestingly, in inflamed tissues, oxygen concentrations are often far below physiological levels [11]. Hypoxia results in adaptive changes to the transcriptions of a wide range of genes involved in increasing oxygen availability to tissues and decreases the cellular consumption of oxygen such as angiogenesis or glycolysis or apoptosis via HIFs-α [10]. Differential activity and expression pattern of HIF-1α and HIF-2α in the tissues have been well identified [12]. In cell cycle progression, stabilization of HIF-1α results in cell cycle arrest due to the inhibition of the transcriptional activity of c-Myc. On the other hand, HIF-2α exhibits opposing effects that is, it increases cell proliferation by activating c-Myc [13].
Many reports have investigated that hypoxia attenuates the expressions of tumor suppressor genes, such as, p53, von Hippel Lindau (VHL), MLH1, BRCA1, 2, and RUNX3 including RECK in normal and cancer cells [8,9,[14][15][16][17]. The products of these tumor suppressor genes primarily function as gatekeepers of cell proliferation, and thus, loss of function or silencing of the tumor suppressors play critical roles in the processes that permit the unregulated proliferation and transformation of normal cells under precancerous hypoxic conditions [18][19][20]. However, biological functions and significances of RECK silencing under hypoxic conditions in hyperplastic phenotypes of early tumorigenesis have largely unknown.
Here we demonstrate that RECK silencing under hypoxic conditions induced cell proliferation in normal epithelial cells and their tumorigenic potential such as anchorage-independent colony forming ability. Knock-down of RECK expression by siRNAs increased c-Myc-mediated cell cycle progression. Our results also suggest that RECK might be an upstream regulator that suppresses HIF-2α through EGFR in obtaining an early tumorigenic hyperplastic phenotype. Therefore, our data reveal a novel mechanism of RECK and hypoxic conditions for the induction of hyperplastic cells in an early step of tumorigenesis through HIF-2α and EGFR.

Ethnics Statement and Chemicals
Animal care and experimentation were performed in accordance with procedures approved by the KNU (Kyngpook National Unversity) Animal Care and Use Committee. An inhibitor for MMP-2 and -9 (#444241), an ERK MAPK inhibitor, PD98059 were purchased from Merck Millipore, and gefitinib, an EGFR specific inhibitor was purchased from Cayman Chemical (Ann Arbor, MD).

Semiquantitative and real-time reverse transcription (RT)-PCR
Gene expressions were analyzed using semiquantitative and real-time RT-PCR as previously described [21].

siRNA experiment and transfection
Two double strand siRNAs designed to target RECK (Qiagen, Valencia, CA), HIF-1α, and HIF-2α and a scrambled siRNA were synthesized (Bioneer, Daejeon, Korea). HEK293 cells were transfected with double strand RNA using HiPerFect Transfection reagent (Qiagen) according to the manufacturer's protocol. RECK mRNA expression was evaluated by RT-PCR and RECK protein expression was monitored by western blotting at the indicated times post-transfection. Proliferation assays were performed on 96 well after splitting cells at 24 h post-transfection. SiRNA1 was designed to target 5'-CAGATTGAAGCCTGCAATAAA-3', and siRNA2 was designed to target 5'-ATACCTGTTCTTGATATTAAA-3'. pBluescript RECK full-length plasmid (Invitrogen) was cloned into pCMV5 vector after digestion with KpnI and SacI. Plasmids were transfected into HEK293 cells that had been plated at 1x10 6 cells per 60 mm dish one day previously. Transfections were performed using Lipofectamine 2000 reagent (Invitrogen) according to the manufacturer's instructions. At 24 h posttransfection, cells were plated onto 96-well plates for proliferation assays or onto 60 mm dishes for protein collection.

Western blot analysis
Equal amounts of protein extracts in SDS-lysis buffer were subjected to SDS-PAGE analysis and then transferred to nitrocellulose membranes. We used the following human antibodies: anti-RECK (BD Bioscience, San Diego, CA), cyclin D1, cyclin A, p27, NFkB (Santa Cruz Biotechnology, Santa Cruz, CA), p21 (Cell signaling), phosphorylated EGFR, ERK1/2 (Cell Signaling), EGFR (Santa Cruz Biotechnology, Santa Cruz, CA), β-actin, and α-tubulin (Sigma, St Louis, MO). An enhanced chemiluminescence system (Pierce, Woburn, MA) was used for detection. Antibody for pEGFR may cross-reacts slightly with other EGFR family member (eg. ErbB2) or activated PDGF receptor. Cells (5 x 10 3 ) were seeded onto a 96-well plate. After 24 hours (time 0), we started to measure cell proliferation at the indicated time points. When we seeded 5 x 10 3 cells in wells, one day later densities became around 20%. Because the cell proliferation rate in hypoxia was higher than in normoxia, only hypoxic cells were almost confluent (-95%) after 72 h of incubation. The graph represents fold changes in numbers of cells at each time versus control cell number at time 0. Cells were plated at 1x10 5 per 60 mm dish and exposed to normoxic or hypoxic condition for the indicated times. The DNA contents and cell cycle statuses of permeabilized cells were determined by propidium iodide (PI, Sigma-Aldrich) staining. Stained cells were processed using a fluorescence-activated cell sorter (FACS, Coulter Elite ESP Cell Sorter, Beckman) and cell cycle profiles were analyzed.

Soft agar colony formation assay
Soft agar colony formation assays were performed using the CytoSelect 96-well Cell Transformation Assay kit (Cell Biolabs, San Diego, CA). Ten days after seeding, the numbers and morphologies of colonies were determined using an inverted phase-contrast microscope (Olympus, Japan). To quantify anchorage-independent growth, colonies were lysed with lysis buffer and detected with CyQuant GR dye. Fluorescence was measured using a fluorometer and a 485/520 filter set (Wallac Victor3 1420 mutilabel counter, PE). Data are presented as the means ± SDs of three independent wells.

RECK shRNA stable transfection
To establish RECK shRNA stable transfectants, we synthesized shRECK oligonucleotides harboring the RECK siRNA1 sequence and cloned them into a blasticidine-resistant pBLOCK-iT Gateway vector (Invitrogen), according to the manufacturer's instructions. For a control vector, we used a lamin shRNA entry clone and used the same procedure used for the cloning of RECK shRNA into the destination vector. shRNA transfectants were selected by treating cells with blasticidine (5 µg/ml) for 21 days. Cells stably transfected with shRECK or shlamin were established and treated weekly with blasticidine (5 µg/ml) in DMEM containing 10% FBS.

Animal experiments and immunohistochemistry
Balb/c-nu mice were purchased from the Institute of Medical Science, University of Tokyo (Tokyo) and maintained in a specific pathogen free facility in accordance with the guidelines issued by the Animal Care and Use Committee of Kyngpook National University. Animals were provided with autoclaved tap water and lab chow ad libitum and were housed in a 23 ± 0.5°C, 55 ± 10 % relative humidity environment under a 12 hour light-dark cycle. For tumorigenesis experiments, 1x10 7 cells stably transfected with shRECK or shlamin were suspended in 0.2 ml of Matrigel and injected into both flanks of nude mice. Tumor growths were determined by measuring the dimensions of tumors with a caliper every two days and multiplying tumor height x length x depth (mm 3 ). Tumor tissues were fixed in 4% paraformaldehyde (pH 7.4) and embedded in paraffin or OCT. Serial sections (5 µm) were mounted on poly-L-lysine coated slides, and processed for either histology or immunohistochemistry. Sections were immunostained with antibodies against mouse CD31 (BD, NJ), p-pRb, PCNA, p16, CA9, RECK (Cell Signaling), and a hypoxic probe (Chemicon International), and visualized using appropriate biotinconjugated secondary antibodies followed by immunoperoxidase detection (Vectastain ABC Elite kit; Linaris, Germany) using diamino-benzidine (DAB) as substrate (Vector, U.K). Counterstaining was performed with hematoxylin. Hypoxic probe-1 (pimonidazol HCl, Chemicon International) was injected 4 h before tumor excision.

Statistical analysis
ANOVA was performed to assess the significances of differences between experimental groups. Statistical significance was accepted for p values < 0.01. Results are represented as means ± SD.

Hypoxia stimulated proliferations of human epithelial cells
To identify whether the downregulation of RECK is involved in regulation of cellular growth or not, we blocked RECK expression using siRNA and performed cell proliferation assays in HEK293 kidney epithelial cells under normoxic conditions. RECK mRNA and protein levels were clearly diminished at 48 h after transfection with two different siRNAs ( Figure 1A). Downregulation of RECK by siRNAs clearly increased cell growth with time, and cell growth rate was inversely correlated with the degree of RECK suppression by siRNA 1 (by 3.5 fold) or 2 (by 2.9 fold) ( Figure 1B).
Previously, we found that RECK was silenced at the transcriptional level by HDAC and HIF-1α in HEK 293 cells under hypoxic conditions [8]. When we examined RECK protein levels in four types of normal epithelial cells isolated from specific tissues, HEK293 (human kidney), PrEC (human prostate), TCMK (mouse kidney), and MCF10A (human breast) after exposed to hypoxia, we observed RECK expression was decreased in a time-dependent manner under hypoxic conditions ( Figure 1C). Exposure to hypoxia for 6 h significantly stimulated the proliferations of HEK293, PrEC, TCMK, and MCF10A cells. As compared with cells maintained under normoxic conditions, proliferations increased by 2.0-3.2 fold after 12-16 h of hypoxia and by 2.3-4.0 fold after 24 h ( Figure  1D), when proliferations maintained a steady state due to confluent cell densities. Because the cell proliferation rate in hypoxia was higher than in normoxia, only hypoxic cells were almost confluent (-95%) after 72 h of incubation. We restored RECK expression with a full-length RECK-pCMV5 plasmid, and then found that cell proliferation was significantly inhibited under hypoxia ( Figure 1E). These results strongly suggest that hypoxia-induced RECK downregulation is responsible for obtaining hyperplastic properties of normal epithelial cells.

RECK downregulation induced by hypoxia was responsible for cell proliferation
We next measured the levels of various cell cycle proteins under hypoxic conditions and in parallel with the silencing of RECK in HEK293 cells. Levels of cyclin D1, cyclin A, and phosphorylated-Rb (p-pRb), as well as c-Myc (a critical factor for G1-S transition) were increased both by hypoxic conditions and by the silencing of RECK (Figure 2A, B). Consistently, levels of p21 cip1 , p27 kip1 and p16ink4A that inhibit cyclin/CDK activity at the G1-S transition point, were depressed under hypoxic conditions and by the silencing of RECK (Figure 2A, B) and these results are in-line with our observations of increased cell proliferation shown in Figure 1. These results can suggest that hypoxia-induced RECK downregulation is responsible for obtaining hyperplastic properties of normal epithelial cells. To determine if hypoxia-induced cell proliferation was due to a decrease in cell death and/or an increase in cell cycle  Figure  2C, see the dashed area for diploid S phase in 12 h hypoxia). Cell death was monitored by counting subG1 phase populations, and the measurements obtained suggested that hypoxia did not induce cell death ( Figure 2C). These results suggest that an increase the S-phase transition and a lack of change in cell death collaborate to stimulate cell viability and proliferation under hypoxic conditions.

HIF-2α was involved in proliferation of epithelial cells under hypoxic conditions induced by RECK silencing
Because HIF-1α and HIF-2α differentially regulate cell cycle progression under hypoxia, we checked the expressions of HIF-1α and HIF-2α under hypoxic and RECK-knockdown conditions, hypoxia increased HIF-1α and HIF-2α protein levels. However, although RECK knockdown induced the upregulation of HIF-2α expression, it did not upregulate HIF-1α ( Figure 3A), which suggested RECK is an upstream regulator of HIF-2α but not of HIF-1α. Surprisingly, NFκB and phosphorylated STAT1 was also upregulated by RECK knockdown and hypoxia ( Figure 3A). HIF-2α mRNA was upregulated by RECK knockdown and hypoxia ( Figure 3B), suggesting RECK regulates HIF-2α at the transcriptional level.
Hypoxia-induced cell proliferation was blocked in HIF-1α and/or HIF-2α silenced cells from 6 h to 24 h of culture under hypoxic conditions ( Figure 3C). We found HIF-1α and HIF-2α siRNAs were effective and confirmed that silencing of HIF-1α inhibited RECK mRNA expression but silencing of HIF-2α did not. Furthermore, the mRNA expressions of HIF-1α and/or HIF-2α under hypoxic conditions were unchanged ( Figure 3C, lower panel).
After 24 h of culture under hypoxic conditions, hypoxiainduced cell proliferation was blocked in HIF-1α and/or HIF-2α silencing cells ( Figure 3C). The results may explain that HIF-2α is involved in the cell proliferation mediated by RECK-silencing whereas HIF-1α is directly involved in RECK gene silencing at the transcriptional level [8].

Involvement of EGFR signaling in the HIF-2α expression induced by RECK silencing
The fact that RECK inhibits the EGFR signaling pathway [5] made us to examine whether the inhibition of EGFR influences HIF-2α expression induced by RECK silencing. After we confirmed that HIF-2α was upregulated under hypoxia in a time-dependent manner ( Figure 4A, upper panel), we treated siRECK-transfected or control cells with gefitinib (Ge, a selective EGFR inhibitor) and PD98059 (PD) under normoxia or hypoxia. Gefitinib suppressed RECK silencing-or hypoxiainduced HIF-2α expression, and PD98059 (a MEK inhibitor) had the same effect as gefitinib on HIF-2α expression ( Figure  4A, lower panel). To confirm this result, we checked the activation of EGFR and of its downstream molecule, p42/44 ERK MAPK, which is activated and involved in RECK silencing in hypoxia [21], and found that hypoxia and RECK silencing both activated EGFR and p42/44 ERK MAPK ( Figure 4B). In addition, we further investigated whether EGFR activation is regulated by MMP or by the silencing of HIF-1α or HIF-2α. Levels of phosphorylated EGFR and downstream phosphorylated ERK were unchanged by MMP inhibitor (Mib) or siHIF-2α under hypoxia, but changed by siHIF-1α ( Figure  4C), indicating the pathway responsible for MMP inhibition differs from required for EGFR activation, and that EGFR activation by RECK silencing is a downstream of HIF-1α and upstream of HIF-2α. These results indicate that the expression of HIF-2α is induced by silencing of RECK via activation of the EGFR signaling pathway.

MMP activation was involved in hypoxia-stimulated proliferation
RECK functions as an inhibitor of MMP activation [22,23]. Therefore, we examined MMPs and MT1-MMP by zymography and western blotting under hypoxic conditions. Hypoxia and siRECK transfection significantly induced pro-and active MMP-9 and MMP-2 by zymography, and increased MT1-MMP protein expression levels ( Figure 5A).
Cell proliferation was increased by hypoxia and by RECK siRNA transfection but treatment with the MMP inhibitor decreased cell proliferation induced by hypoxia or siRECK transfection ( Figure 5B). These findings suggest that MMP-2 and MMP-9 are involved in the activation of cell proliferation induced by the silencing of RECK under hypoxic conditions, but that the pathway involves differs from that of EGFR-HIF-2αinduced cell proliferation.

Silencing of RECK gene promoted in vitro and in vivo tumor formation
To confirm that the downregulation of RECK promotes the hyperplastic activity of epithelial cells in vitro, we used soft agar colony formation assays [24]. Cells transfected with RECK siRNA formed more colonies than those transfected with scrambled siRNA (Figure 6A, a & c vs. b & d), and the fluoroactivities of colony lysates from RECK siRNA transfected cells were significantly higher than those of control siRNA transfectants ( Figure 6A, graph).
To explore the role of RECK downregulation in hyperplastic development during early tumorigenesis, we generated stable cells with RECK-knockdown. These cells were then injected s.c. into nude mice flank, and tumor masses were measured from 7 days post-injection for 22 days. As early as one week post-injection, nude mice injected with shRECK transfectants bore tumors with greater mass volumes and growth compared to those injected with control cells ( Figure 6B). Tumor microvessel density assayed by CD31 immunostaining showed that RECK knockdown was associated with highly vascularized tumors ( Figure 6C, a-b). Furthermore, levels of p-pRb and of proliferating cell nuclear antigen (PCNA) were markedly higher in shRECK tumors, but p16 INK4A expression was completely abolished ( Figure 6C, c-h). To confirm the inverse relation between hypoxia and RECK expression in tumors, we examined hypoxic regions in RECK silenced tumors by injecting pimonidazole (a hypoxic probe). Immunohistochemistry showed that hypoxic regions ( Figure  5D) and carbonic anhydrase IX (CA9) (a hypoxic marker protein) expressions were greater in RECK silenced tumors ( Figure 6D). Taken together, these results suggest that RECK suppression by hypoxia promotes the acquisition of a premalignant hyperplastic phenotype and enhances in vitro and in vivo tumorigenesis via G1/S phase cell cycle progression through the activation of EGFR and HIF-2α.

Discussion
In this study, RECK expression was found to be downregulated by hypoxia in normal epithelial cells and to be involved in the development of the hypoxia-induced hyperplastic phenotype. RECK silencing under hypoxic conditions enhanced cell cycle progression through c-Myc, and silencing of RECK induced EGFR activation and subsequently HIF-2α expression. Colony formation assays and in vivo tumor xenograft experiments suggested that the suppression of RECK by hypoxic stress is required to transform normal cells to tumor-like cells by enhancing their proliferative abilities. Therefore, loss of RECK expression by microenvironmental hypoxic stress may stimulate to get a more hyperplastic phenotype in normal epithelial cells and to favor early tumorigenesis events. Our results support previous suggestions that RECK functions as a transformation suppressor gene in premalignant and normal tissues [3,23].
We found although hypoxia stabilized HIF-1α and HIF-2α, only HIF-2α was upregulated by silencing of RECK, and that knockdown of either HIF-1α or HIF-2α recovered hypoxiainduced cell proliferation (Figure 3). Because HIF-1α participates in RECK downregulation at the transcription level [8], we can speculate that HIF-1α and HIF-2α are upstream and downstream regulators of RECK, respectively under hypoxic conditions. But HIF-2α is directly involved in RECKsilencing-mediated cell proliferation.
HIF-2α and c-Myc may collaborate to increase cell proliferation under hypoxic conditions and in RECK silencing cells. HIF-1α and HIF-2α have opposing effects on c-Myc interaction and cell proliferation [12]. Under hypoxic conditions, HIF-1α and HIF-2α are competent to occupy c-Myc. HIF-1α participates in gene silencing at RECK promoter in accompany with HDAC1 [8], and in the donating of c-Myc to HIF-2α rather than binding to c-Myc to cause hypoxia-induced cell proliferation. Since HIF-1α and -2α are expressed in a cell-type specific manner [25], hypoxic conditions might have different effects on cell proliferation depending on the availabilities of HIF-1α and HIF-2α in a variety of cells. In a recent study, HIF-2α was present in a stem cell population but HIF-1α was present in both stem and non-stem cell populations and was only stabilized under more severe hypoxic conditions [26]. HIF-2α transcriptionally activates the Oct4 gene, a stem cell marker [27]. A significant evidence supports our hypothesis that HIF-2α alters the basic genetic activity of normal transformed cells to a more stem-like hyperplastic phenotype in a manner that depends on the fluctuating state of oxygen availability under chronic pathophysiological conditions [28]. Furthermore, a switch of hypoxic response from HIF-1 to HIF-2-depedent gene expression results in high aggressive tumor through the promotion of stem cell characteristics [29]. In CD133+ glioblastoma cells, the expressions of MMPs and RECK were induced by miR-125b, and the authors suggested that the functional inhibition of RECK increases cancer stem cell function [30]. Our findings suggest that RECK is likely to be an important molecule for regulation of switch from HIF-1α-to HIF-2α-dependent transcription.
Although it has been shown that hypoxia induces cell cycle arrest but not proliferation [31,32], exposure time and degree of hypoxia can also affect cell cycle arrest/survival decisions. For example, longer and more severe hypoxia induces cell cycle arrest and apoptosis to a greater extent than shorter and milder exposure [32]. Hypoxia and the silencing of RECK by siRNA  increased the protein expressions of c-Myc, cyclin D1, cyclin A, and p-pRb, but decreased those of p21 cip1 , p27 kip1 and p16 ink4A , indicating that the G1/S transition is a key regulation point of the hypoxia-induced proliferation of epithelial cells. These results are consistent with the previous findings that RECK depletion stimulates cell proliferation during successive cell cultivations, and that this is correlated with diminished expressions of p21 cip1 , p53, and p19 ink4D in mouse MEFs [5].
In cancer cells, the repression of RECK by hypoxia also influenced cell proliferation and enhanced migration or invasion [21,23,33]. Yoshida et al. found that the inhibition of cancer cell proliferation is mediated by S-phase kinase-associated protein 2 (SKP2) targeting p27 kip1 [34], which suggests that different molecules might be involved in the suppression of cell growth by RECK in a cell-specific manner. Thus, the downregulation of RECK and the stabilization of HIF-2α by hypoxia may contribute to the development of the hyperplastic phenotype in premalignant cells by increasing the expressions of genes involved in cell cycle progression. The activation of HIF-2α in cancer cells increases the expression of EGFR [35] in response to hypoxia [36]. In a study using EGFR inhibitor, it was suggested that HIF-1α might be a downstream regulator of EGFR [37]. However, our findings suggest that EGFR might be involved in the upregulation of HIF-2α in RECK-silenced cells and in cells under hypoxic conditions ( Figure 4B). Importantly, findings in VHL-/-cancer cells demonstrated that the shRNAmediated inhibition of EGFR was sufficient to abolish HIF-2αinduced spheroid formation in a three-dimensional tumorigenic assay [38], which suggests that EGFR is required for HIF-2αmediated tumorigenesis. Furthermore, the inactivation of EGFR can be achieved by the restoration of RECK [5]. Therefore, it is suggested that the RECK-EGFR-HIF-2α axis is a critical determinant of the hyperplastic phenotype in hypoxic cells.
Because an MMP inhibitor decreased proliferation rate induced by hypoxia, MMP inhibition by hypoxia-induced RECK silencing involves a different pathway from involving EGFR-HIF-2α during the inhibition of cell proliferation. As MMPs are also involved in cell proliferation [39], MMP inhibition by RECK downregulation also suppresses cell proliferation induced by hypoxia. These findings suggest that these two pathways work together to increase cell proliferation.
Interestingly, it has been previously reported that the activation of NFκB in intestinal epithelial cells and macrophages results in the inductions of anti-apoptotic genes and cytokines that increase the survival and proliferation of premalignant intestinal epithelial cells [40]. In our findings, activation of NFκB and STAT1 both by hypoxia and RECK silencing may also participate the cell proliferation and survival.
The restoration of RECK and the destabilization of HIF-2α could perhaps be adopted as a means of preventing the early transformation of hyperplastic cells in premalignant lesions under hypoxic conditions. However, further investigations are needed to characterize the mechanism by which RECK regulates c-Myc or HIF-2α during transformation to the hyperplastic phenotype and during the cell cycle progression of normal cells. The observation that hypoxia can cause RECK silencing in normal cells is important, as it could be a novel mechanism for reducing tumor suppressor expression under low oxygen tension-dependent stress, which characteristic of certain pathological conditions.